In the United States today, most residential heating systems are single zone. This means that all rooms must be either conditioned or not, which often leads to large unoccupied areas being conditioned to satisfy comfort needs in the occupied zones. Furthermore, these systems typically have only one sensor, which is often not representative of all zones. It would be desirable to have better control, but multizone systems are expensive and invasive to install if one wants to retrofit an existing residential single zone heating system into a multizone system.
A new multizone system is currently being developed at UC Berkeley to provide an inexpensive, simple retrofit strategy. This consists of a wireless vent register control system. The new registers have been fitted with a mote (a small wireless device) that can send and receive signals in order to open and close vents in a house, allowing the system to only condition the desired zones. Motes fitted with temperature sensors are also placed in the zones and outside to provide additional temperature readings to the control system. Therefore, the house can take on properties of a multizone system by only changing the registers. The goal of the overall project is to design a simple, inexpensive system that can save energy while providing thermal comfort comparable to or better than that of a traditional single zone system.
Since smaller volumes will be conditioned at a time, the heating system can potentially reach setpoint quicker, therefore running for a shorter period of time. The new system has been tested using a 1988 house in Danville, CA. This test was to verify that the system works as expected and whether target zones reach setpoint faster. While multizone systems can conceptually allow the user to fine-tune conditioning needs, the control algorithms will be more complicated. A single zone system only needs to actuate heating, cooling, and fan modes with one setpoint at a time. With the wireless vent control, the system must track which zones need to be conditioned and choose an appropriate vent configuration, in addition to the control needs of the single zone system. To optimize this system in a real house would be prohibitively complicated and time consuming. Therefore, an optimization model has been developed to determine optimal control sequences based on setpoint and occupancy. This optimization models the actual house, using real house data such as insulation values, flow rates, and fan power. It automates the choice of optimal configuration, rather than completing thousands and thousands of runs in real time in the actual house.
The Background section of this thesis will provide an overview of existing residential heating options. Specifically, it outlines the research that has been done on the effectiveness of existing multizone control strategies. It will also describe the new multizone system being developed in the UC Berkeley Demand Response Enabling Technology group. This system provides wireless control of vent registers. A statement of the problem and objectives follows the background.
Next the Methods section provides a detailed description of the hardware and software used in the wireless multizone control system. The development of the hardware and software for this project is the subject of William Watt’s thesis, Multizone HVAC Control with Smart Vent Louvers (2007). This section discusses the initial house testing and the measurements. The Methods section also contains an overview of the optimization model developed in this work, including an explanation of the thermal calculations, logic, occupancy schedules, and weather data. The initial optimization model (and the resulting write-up) began as a final project in ER 220: Modeling Energy, Environmental, and Resource Systems. Petek Gursel and Arman Shehabi contributed to this project.
This thesis contains an updated model with enhanced thermal calculations and more detailed analysis. The Results section provides tables and figures from both the basecase (single zone control) and the multizone control runs, and also presents the sensitivity analysis.
The Discussion, Future Work, Conclusion, and References sections follow. Finally, many people have been involved in the completion of this thesis. Acknowledgements can be found at the end of the paper.